When we approach modern Machine Learning problems in an AWS environment, there is more than traditional data preparation, model training, and final inferences to consider. Also, pure computing power is not the only concern we must deal with in creating an ML solution. There is a substantial difference between creating and testing a Machine Learning model inside a Jupyter Notebook locally and releasing it on a production infrastructure capable of generating business value. The complexities of going live with a Machine Learning workflow in the Cloud are called a deployment gap and we will see together through this article how to tackle it by combining speed and agility in modeling and training with criteria of solidity, scalability, and resilience required by production environments. The procedure we'll dive into is similar to what happened with the DevOps model for "traditional" software development, and the MLOps paradigm, this is how we call it, is commonly proposed as "an end-to-end process to design, create and manage Machine Learning applications in a reproducible, testable and evolutionary way". So as we will guide you through the following paragraphs, we will dive deep into the reasons and principles behind the MLOps paradigm and how it easily relates to the AWS ecosystem and the best practices of the AWS Well-Architected Framework. As said before, Machine Learning workloads can be essentially seen as complex pieces of software, so we can still apply "traditional" software practices.
Oct-5-2021, 16:02:43 GMT